Three years after igniting the AI revolution, OpenAI is "forced" to go public.

In November 2022, ChatGPT was launched, and its servers were overwhelmed on the very first day. Prior to this, no product from any technology company had ever acquired 100 million users at such an astonishing speed.

After that moment, the "AI revolution" transformed from an industry buzzword into a reality that ordinary people could feel.

It was OpenAI that kicked off this revolution.

Three years later, according to media reports such as The Information, OpenAI is preparing to secretly submit a draft IPO to the SEC and aims to go public as early as September, with a target valuation of over $1 trillion. Altman didn't deny this report.

But just 48 hours ago, he told employees at an all - hands meeting that submitting an IPO application and being truly ready to go public are two different things, and the company won't rush to the public market before the conditions are ripe.

These two statements clearly illustrate OpenAI's current situation.

Forced into action

Altman said it's not ready, but the company is reported to submit the form today. To understand this contradiction, we need to first look at the three - fold pressure behind it.

First, Anthropic is overtaking.

In the past 15 months, Anthropic's annualized revenue has grown from $1 billion to $30 billion, a 30 - fold increase. During the same period, OpenAI's revenue increased from about $20 billion to $25 billion, a rise of about 25%.

In April this year, Anthropic's annualized revenue officially exceeded that of OpenAI.

The uglier figures are in Q2. Anthropic expects its Q2 revenue to be about $10.9 billion and will achieve an operating profit of about $600 million. Anthropic will turn profitable before OpenAI.

Anthropic's private equity valuation is already $1.2 trillion, higher than OpenAI's self - set listing target of $1 trillion.

And Anthropic is also preparing for an IPO, as early as October. Altman privately said that he hopes OpenAI will be the first to go public.

Second, SpaceX will be seeking funds next month.

SpaceX, owned by Musk, plans to raise about $75 billion at a valuation of $1.75 trillion next month. This will be the largest IPO in history.

According to reports citing insiders, there is "a lot of competition" between OpenAI and SpaceX. OpenAI hopes investors will set aside funds for it in advance and not put all their money into SpaceX.

The competition between the two companies for public market funds is real.

Musk was once a co - founder of OpenAI. He later left and founded the rival xAI, and the two companies are still in a lawsuit.

Now, another one of Musk's companies is going to compete for funds from the same pool.

Third, the demand for computing power is insatiable, and there isn't enough money.

OpenAI expects to spend $665 billion on computing power before 2030. The company has raised nearly $200 billion in total so far, but it's still a long way from that figure.

Going public to raise funds is the next opportunity to replenish resources, and there are no alternative options.

With these three - fold pressures combined, the logic is clear: It's not that OpenAI is ready, but that it can't afford to wait.

Lack of internal unity

For this IPO, the company is far from being in sync internally.

Large investors are surprised by the progress speed, which is not a normal sign before an IPO.

OpenAI's CFO, Sarah Friar, is significantly more cautious about the listing pace than Altman. This was specifically reported, indicating that there are real differences within the company.

Altman himself mentioned at the all - hands meeting that SpaceX's upcoming large - scale IPO and the global economic trend will be the key external variables affecting whether OpenAI can finally go public. In other words, he himself isn't sure if the window in September will open.

This doesn't seem like the state of a company confident in its IPO. It's more like: the situation is forcing it forward, but even the insiders aren't sure how far it can go.

Cracks in the story

A deeper problem is that the story OpenAI tells to the public market is becoming increasingly difficult to justify.

ChatGPT became an instant hit back then, establishing OpenAI's unshakable brand advantage in the consumer market. However, in the past 12 months, ChatGPT's web traffic market share has dropped from 87% to 68%, while Google Gemini's has risen from 5% to over 18%, and its monthly active users have increased from 350 million to 750 million.

Google has search, Android, and Chrome - these are entry points that OpenAI doesn't have and can't buy.

The enterprise market is also under pressure. In direct head - to - head competition, Claude won about 70% of enterprise orders.

Anthropic, with fewer users (134 million vs. OpenAI's approximately 900 million), has achieved a higher market revenue share (31.4% vs. 29%), and the average monthly revenue per active user is $16.2, leading in business efficiency.

These losses in both directions have occurred in OpenAI's core market.

OpenAI's response is to bet on the AI Agent economy. If AI Agents become the main way for humans to interact with machines, OpenAI's models will be the operating system of this economy, and the $280 billion in revenue in 2030 will just be a small part of the ecosystem.

This logic holds at the conceptual level. But it requires OpenAI to maintain growth in the consumer market, the enterprise market, and the developer ecosystem simultaneously. However, the competitors in all three battlefields are accelerating, while OpenAI's own growth rate is only 25%.

Going from $25 billion to $280 billion in 4 years, with a 65% compound annual growth rate - this is the figure that the public market will examine.

The most expensive self - reflection

This is the first time OpenAI has to disclose its financial statements.

Previously, the $852 billion private equity valuation was based on the shared narrative of a few institutional investors. The financial data was opaque, and the valuation methods were inconsistent. Once the S - 1 is made public, the loss amount, cost structure, Microsoft dependence terms, and the specific assumptions of the $280 billion forecast will all be exposed to the public market.

The questions from the public market are straightforward: Can OpenAI win? When will it turn profitable? Why is it worth this price?

This is exactly what Altman meant when he said "the conditions aren't ripe yet." He knows better than anyone that the private market buys the story, while the public market buys the numbers. And OpenAI's current numbers - slowing revenue growth, declining market share, and still in the red - aren't the best situation for going public.

But it no longer has the luxury of waiting for the best situation.

With each passing quarter, the gap with Anthropic narrows. With each passing day of SpaceX's progress, there is less money in the market. The computing power bill increases every month.

This is what it means to be on a tiger: it's easy to get on, but hard to get off.

OpenAI has turned itself into the biggest symbol of this AI revolution in three years. Now, this symbol itself has become a burden - it has to live up to the $1 trillion valuation, even if it's not ready.

Asset impact

Microsoft is the biggest beneficiary of this IPO. With an investment of $13.8 billion for a 27% stake, if the IPO is priced at $1 trillion, the floating profit will exceed $270 billion. It's positive in the short term, and the investment logic is clear.

However, it should be noted that after OpenAI goes public, it will seek to reduce its dependence on Microsoft, and Microsoft's exclusive AI advantage may be diluted in the medium to long term.

Nvidia: OpenAI is one of the largest buyers of computing power. Going public to raise funds → accelerating computing power expenditure → supporting GPU demand. But in the medium to long term, OpenAI has a roadmap for self - developed chips, which is a variable that needs to be discounted in the valuation.

CoreWeave: As an OpenAI computing power cooperation partner, it will directly benefit from the acceleration of computing power expenditure and can be used as a proxy target.

AI sector in A - shares: If the $1 trillion pricing is successful, it will positively anchor the global AI valuation; if the IPO is priced at a discount or postponed, the AI bubble narrative will face phased pressure.

Subsequent tracking

SpaceX IPO (next month): Whether it will draw funds from the public market and whether OpenAI's roadshow schedule will be forced to be adjusted.

Anthropic Q2 performance: If the $10.9 billion in revenue and profitability are realized, it will directly suppress the narrative before OpenAI's roadshow.

Anthropic IPO timing (as early as October): Which company goes public first and which one gets a higher valuation will be the most important pricing event for the AI sector this year.

Public version of S - 1: Any figure that exceeds expectations, such as the loss amount, Microsoft contract terms, and the breakdown of the $280 billion forecast assumptions, will trigger a market re - pricing.

Whether Altman changes his timing judgment: He said that market conditions are the ultimate variable.

Altman brought ChatGPT to the world and initiated this AI revolution. But the revolution won't wait for someone to get themselves ready.

Now, he's sitting on a running tiger's back. There are Anthropic, SpaceX, and Google in front, and a bottomless computing power bill is following behind.

It's hard to get off. It's also hard to stay on.

This article is from the WeChat official account "Wall Street Insights", author: Insights Research Group, published by 36Kr with authorization.